Graph Theory in High School Education

D. Lessner
Department of Software and Computer Science Education, Faculty of Mathematics and
Physics, Charles University, Prague, Czech Republic.

Abstract. Graph theory is not being taught on Czech high schools. The reasons
seem to be rather historical than factual. In this contribution, we discuss the
usefulness of elementary Graph theory for High school students. Then we formulate
the targets and objectives of this education. Core of this contribution is topological
sorting. It is an example of an advanced topic in graph theory, which is still
comprehensible for high school students. Moreover, it is potentially useful in
their life also out of school. We explain the problem, its applications, and three
approaches to the solution along with comments regarding introduction of such
topic into high schools. Last part of this contribution explains the critical path
method, another advanced topic strongly bond to the idea of representing tasks as
graphs.

Introduction
Computer science is not a part of high school education1 in Czechia [Lessner, 2010]. We
believe it to be a mistake. Our research addresses this issue. Graph theory is one of the topics
we would like to introduce to our schools. In this contribution, we provide some more details to
this specific topic. It shows our point of view on graphs for the sake of high school education.
First we clarify what we consider to be the basics of graph theory. For high school students
the classical G = (V, E) approach is rather confusing. What they really need to know is that
graphs are structures to represents relations between objects and that it is useful to try to
draw them. Then we examine the situation of graph theory on high schools. Graphs are not
included in the Framework Education Programme (fundamental curricular document for our
high schools, [VÚP, 2007]) nor other programmes ([CERMAT, 2010]) explicitly, yet they are
used in many areas intuitively. Sadly, there is no awareness of some unifying theory which
would allow knowledge transfer between different areas using graphs. After explaining some
more reasons why graph theory should be introduced to high school students we describe the
main educational targets.
After these preparations, we describe a more advanced topic in closer detail. We have
chosen topological sorting for this purpose. The problem is to find an order of vertices in
oriented graph which would respect the edges orientation. The usual idea is that vertices are
tasks and edges their dependencies. We will show three approaches to solve this problem and
discuss the interesting points from the high school point of view. Developing the problem of
topological sorting into an even more useful idea we introduce the notion critical path, as well
as a way how to find it and use it. At the end we summarize the whole contribution.

Graph theory on high schools

In this section we discuss graph theory education on high school level in Czechia. At first
we briefly revisit graph theory itself. Then we look into the recent situation on out high schools
regarding graph theory. Then we can discuss the reasons why to teach about it as well as
specific educational targets.
Here we revisit the basics about graph theory. At first we shall make clear that graph theory
in computer science is not about graphs of functions or charts. Graph is an abstract structure

1
Secondary education, i. e. 15 – 19 years old in their 10. – 13. year of study.
 LESSNER, D.: GRAPH THEORY IN HIGH SCHOOL EDUCATION

to represent relations between pairs of objects. This is very general and leads to extreme
applicability in a wide range of areas. Among the advantages we see also visuality. Formally,
graphs are sets of vertices and edges (i. e. pairs of vertices). This would be represented by
some sequence of symbols. However, it is much more useful for general population to represent
graphs (i. e. relations between objects) in form of a drawing. This way one can see much deeper
into the structure of represented relations and find much easier what he seeks.
Among the basic notions to study are graph, vertices and edges of course, including ori-
entation, degree and score. Higher terms are path, cycle, coloring and components. There are
some basic graph types we distinguish, mainly complete graphs, trees and regular graphs. An
important topic is graph representation for the sake of further algorithmic processing, e. g. in
computers, where drawing is not really helpful. Among processing we are interested in vari-
ous types of systematic traversing and searching. These may also lead to solve more advanced
problems, like topological sorting, described later in this contribution.
Application of graphs involves nearly any area which deals with some kind of binary rela-
tions between objects. That is actually almost any area of human activity. Classic application
is in maps an plans, where vertices represent places (e. g. towns) and edges their connections
(e. g. roads). This is a very helpful metaphor used to think about graphs. Another applications
rise from the same idea. We do not connect only places on a map, there are also pipes and
tanks or electric circuitry. More applications are mentioned in the following subsection.

Graph theory in official curricula

Let us see the situation of graph theory on Czech high schools. Shortly, it is not there.
This is surprising considering the usefulness of this topic and situation abroad [Gal-Ezer et
al., 1995]. On the other hand, it is no surprise at all, if we consider the overall situation and
historical development. Last years, maths related education is going through severe reduction
of both hours and topics to teach. Interested teachers and researchers focus rather on saving
what is left than on introducing new topics. This shows also the historical background. Graph
theory has simply never been a part of high school programmes. The field itself, regardless of
how rich and important results, is much younger compared to the classical parts of high school
mathematics. Further, should it be officially included, it might be of no effect at all. Without
introducing CS itself, graph theory would most likely become a part of mathematics. The
most important approach to the structure of mathematical programme in Czechia is respecting
historical development of mathematics. Therefore graphs would come after the other topics. In
practice, the last chapters are being taught usually only in specialized seminars for students who
have chosen mathematics as to be a part of their school leaving exams or might need advanced
mathematics for their university studies. These seminars focus mainly on the most important
chapter of the end of high school level mathematics, which is differential calculus. Next to it,
graph theory would be very likely left for self-study.
Although graph theory is not explicitly mentioned anywhere in the high school curriculum,
we may look for its usage, even in seemingly unrelated subjects. The finding is disturbing.
Graphs are being used widely throughout virtually every school subject on intuitive level. We
will mention only a few blatant and obvious examples. In organic chemistry, students draw
molecular structures. Simple math is used to find formulas which can not exist (given the num-
ber of atoms and their connectivity). In biology, taxonomy and genetics uses trees. Metabolic
network is a graph. Royal family trees are needed in history for some nontrivial cases. Usage
in geography and cartography is clear. In language, syntactic structure of a sentence is repre-
sented using a tree. Introduction to combinatorics is based on routes and intersections. Last
application to mention is one of the recent hits in primary and secondary education in Czechia,
mind mapping. For a computer scientist, mind map is nothing special but a tree modification.
Despite of all these examples students encounter, they are unaware of any unifying theory be-
hind. This limits their ability to transfer any knowledge or skill (e. g. tree characteristics or
 LESSNER, D.: GRAPH THEORY IN HIGH SCHOOL EDUCATION

systematic searching) to some different area.

Reasons for including graphs into high school education

We have described applications of graphs in high school education. Considering how fre-
quent they are, they stand as one of the reasons why to include graph theory into high school
programmes. Here we describe also other reasons for it. Graphs represent a new structure to
work with, different from numbers or other classical structures from high school mathematics.
Graphs seemingly require a different way of thinking. This might provide poor skilled students
with a new chance to catch up in mathematics. Even though the structure differ and it may
be mostly about drawing, the reasoning behind is the same as everywhere else. The same idea
works with algorithms here. They seem to be somehow different, working with vertices and
edges instead of numeric values. Yet, this is no fundamental difference here. Graphic algo-
rithms give the opportunity to notice this and therefore understand better what an algorithm
itself is.
It is important to remark that even though the basics are very intuitive, graph theory
naturally provides advanced topics of various difficulty. Most of them have their important
applications as well as theoretical impact. One of the many examples would be the notion
of planarity. Many students think of graphs as of their planar drawing. This is no mistake,
yet they shall know this approach has its limits. It is crucial to discover that many different
drawings of one graph exist and therefore a drawing is nothing more than a visualisation. It is
also important to experience that there are graphs with no correct drawing at all, yet algorithms
work in the very same way. This is about planarity – a natural and useful notion, yet not trivial
at all. Another such topic is isomorphism of graphs. We believe it to be at the limits of high
school students capabilities.
Concluded, the reasons for teaching graph theory is in their practical usefulness as well as
in their challenging theoretical background.

Education targets
Here we introduce shortly the main targets specific for graph theory education, as we have
developed them so far. They are made compatible with the official high school programme
[VÚP, 2007]. To begin with, students shall recognize the appropriate situations to apply graphs
(i. e. not regarding city plans only), and they shall also apply them. This means they shall be
able to read relations from given drawn graphs as well as denoting them (both by drawing and
symbolically). They shall be able to use some unified terminology. This does not necessarily
mean the official terms, especially in cases they are not natural in our language (as “graph”
itself). The point is smooth communication when using graphs. They should be able to use
drawing and sketching to their advantage, yet understand the unavoidable need to rely on rea-
soning in certain situations. Further processing includes translating notations (natural language
description, drawing, list, matrix) and systematic traversing. This does not mean they shall
know the exact way of depth first search for example. The point is to make them understand
a few key points which will make their approach systematic (i. g. marking open and closed
vertices) and which will enable them reinvent the algorithm if needed.
On higher level, they should be aware of some meta knowledge and heuristics, and be able
to use them appropriately. They should be aware that there are many ways to achieve the same
goals (regarding both graph drawing and its algorithmic processing). This goal however does
not imply these ways themselves to be the same as well. They may differ significantly, mostly
in difficulty and efficiency. Influenced by all the examples in graph theory, students shall know
it is good to work systematically and in phases, virtually every time when it is possible.
Further, they shall understand the importance of preprocessing data and other preparations
preceding work itself. One of the many examples is topological sorting, discussed below in this
contribution. On the same example, they shall learn that a greedy approach may work well.
 LESSNER, D.: GRAPH THEORY IN HIGH SCHOOL EDUCATION

By greedy we mean fixating partial solutions and no turning back (unlike in backtracking).
This usually require some more work with choosing which partial solution to pick. Usually the
least injuring one is found and picked. Still, we would like them that such an argument is not
necessarily correct and they shall rather seek guarantee of a partial solution correctness (and
guaranteed inclusion in a correct and complete solution).
Educational targets described above combine both practical usability of graphs and the-
oretical knowledge. These two shall strengthen each other synergistically. Practice leads to
problems, solved using theoretical results, enabling more practice, and so on. This enhances
students lives, providing them with tools to solve problems (using graphs), and make them
understand the importance of quality theoretical thinking at the same time.

Example: Topological sorting

We have chosen the problem of topological searching as an example for this contribution.
It represents a problem which is understandable, challenging, yet well solvable and widely
applicable. It is well known, therefore we will not dive into exact details. We rather save space
for what is interesting in high school context. Detailed explanations can be found in [Kučera,
1983] and [Töpfer, 2007]. In this sections we introduce the problem and the main motivation
for its solving, three main approaches to find the solution and a few further topics.

The problem
On input we have a directed graph. We search for an order of vertices which would respect
edge orientations. More formally, for a given G = (V, E), we search for a φ : V → ℵ, such
that (v1 , v2 ) ∈ E ⇒ φ(v1 ) < φ(v2 ). The usual motivation for this problem is about tasks
(represented as vertices) and their time dependencies (edges), such as getting dressed, building
a villa or constructing a plane. It is not wise to put on shoes before socks, yet shoes are without
any direct relation to a tie. With a complex project, it is easier to find these dependent pairs
than the overall order of tasks.
Such a motivation is relevant to high school students lives. Not only do they encounter
such problems occasionally, it is also important that the problem setting is comprehensible for
them. Moreover, they may expect such (and more complex) problems in their future.

The solution
We will discuss three main approaches to solve the problem of topological sorting. They
differ in many aspects, including efficiency. The first approach to consider is based on brute
force. That would be an average student’s first choice to examine. Try a few permutations of
vertices and examine whether they satisfy given conditions (partial order implied by the graph).
Obviously, with n! possible orders, this may take very long, and moreover, the student will soon
feel that he examines needless number of possibilities. He might notice some repeating patterns
or even partial solutions. Trying to combine and complete them may lead to a solution, as well
as to a few questions: How can we be sure it is correct? How do we do this systematically with
different input data?
During their work on a specific input, they might discover parts of their solution which are
not going to change anymore. Specifically, they should notice the special characteristics of the
first (or last) task to be done: they have no predecessors (or successors). This may lead to find
the desired order systematically instead of a sequence of trial and errors. Excluding already
ordered vertices from the graph (as well as incidental edges), we may proceed recursively. The
graph is smaller, the idea is the same. Task with no predecessor may be put as the first and this
causes no mistake. This can be found out by a high school student, given there is enough time,
appropriate sample problems and careful guidance by the teacher. From this, the algorithm may
be formulated quite easily. These observations gave us also a necessary condition for existence
 LESSNER, D.: GRAPH THEORY IN HIGH SCHOOL EDUCATION

of any solution. Whenever there is no vertice to be chosen (i. e. with no predecessor), there is
no task to start with and therefore no solution at all.
This approach may still be refined. Searching for the first vertice (with zero predecessors)
is unnecessarily time consuming if don’t think it through and do it with no preparation in every
step. This is useful to be experienced by students: Having a smart idea leading to a solution
undoubtedly, yet not efficiently. At the same time, it is a trap for more advanced students. We
want to find the candidates quickly, so one of the ideas could be to sort the vertices according
to their incoming degree and keep a sorted list. Actually, a more efficient way can be found,
because only the first vertice is needed. Therefore some kind of heap could serve us well.
However, heap still means wasting resources, because we know exactly, what of what degree
value shall be the first vertice. Finally, it is enough to keep a list of candidates (vertice with no
predecessor) and the incoming degree for each vertice. Keeping these two structures up to date
can not be faster (we will always need to do at least one operation per vertice and per each of
its edges).
The last possibility we describe here is unlikely to be found by students, as it is not very
natural. It is based on depth first search (DDS). Therefore it may help to revise DDS. Moreover,
this approach may illustrate one of the key concepts in computer science: try to reuse what
we already know for solving a new situation. In this case however, it is not straightforward.
DDS has to be run on a graph with reversed edges. First it finds vertices with no predecessors
(along edges) of this modified graph. Then DFS is run from these vertices one by one. Order of
closing vertices is the demanded topological order. If an open vertice is encountered anytime,
the original graph contains a cycle and therefore there is no correct topological order. Time
complexity of the last two approaches is proportionate to the number of vertices and edges, if
appropriate data structures are used.

Further topics
The topic of topological sorting bring an excellent opportunity to seek some further matter -
both theory and applications. Orders in general may be introduced, as well as lattices and Hasse
diagrams. The strongest topic would probably be deepening planning and scheduling. There is
an important counterpart to this out of CS. One of the major areas of project management is
basically planning and scheduling without the algorithms. We will mention here one possible
and fruitful topic, the critical path method (CPM). For this we need to consider also time
consumed by individual tasks, so we add it as weights to vertices.
Finding an order for given set of tasks with their dependencies is certainly useful, yet often
insufficient in real life. Tasks can often be accomplished in parallel. This of course does not
mean they can be accomplished all at once, there still are their dependencies. A reasonable
question in this situation is about the minimum working time to accomplish all tasks. Length
of the critical path is the answer. Also, this is a way to define critical path. Looking into it, we
would find out it is the longest possible sequence of vertices connected by edges. Critical path
is a good example that some useless looking problems, such as searching for the longest path in
a graph, may actually be of crucial importance in certain situations.
The total time taken by the project can not be shorter for a simple reason. Each task on
the path has to be accomplished, in a time given by its weight. And it has to be accomplished
after its direct predecessor and before the direct successor, not along any of them. This holds
for all vertices on the path. This shows us the importance of finding all critical paths, not only
the length. Should any task on any critical path be delayed, the whole project will inevitably
end late. We should remark here that all this is not dependent on the topological order possibly
found earlier, CPM only needs the graph itself.
Finding a critical path is surprisingly easy, and therefore suitable for a high school student
with a very little computer science experience. We will find the shortest time a vertice may be
finished for all the vertices, one by one, using the topological ordering. For a vertice with no
 LESSNER, D.: GRAPH THEORY IN HIGH SCHOOL EDUCATION

predecessor, the time needed to finish is obviously only the time for the task itself. Whenever
we take a vertice in topological order, all its predecessors have been processed already, therefore
we know the earliest moment they could be accomplished. The earliest moment to finish the
taken vertice one is obviously the time of its last finished predecessor plus the time for the
vertice itself. This way, we simply run through the graph and end with an answer at hand.
This approach reuses many crucial ideas from topological sorting.
Another way to modify the problem and seek for another solution could be program evalua-
tion and review technique (PERT) or hierarchical tasks using work breakdown structure (WBS).
Of course, classically scientific questions may also be solved in the classroom, like what is the
total number of all topological orders on given graph.

Summary
In this contribution, we have proposed a new topic for high school education – graph theory.
We have examined the current situation, shown why we consider it important and what should
be the benefits of including it. Graphs are a very helpful tool for dealing with relations between
pairs of objects. This is the way we want high schools students to see graphs. They shall be able
to use graphs this way, including some basic traversing and searching. After this we discussed
topological sorting in detail. It represents an advanced topic, yet comprehensible for students
and moreover, very useful. Many important aspects of graph theory and computer science may
be illustrated using this problem. In the last part of this contribution we showed a possible
continuation of topological sorting. The critical path method is again a very useful concept, yet
comprehensible by students and using some elegant ideas and reasoning.
High school students are already using simple graphs, intuitively. Yet, they are not aware
of any special science behind. We believe that showing it to them would improve their lives.
They would be able to transfer and use their graph related knowledge and skills in new areas.
They would also understand better the limits of drawings and common sense based approach.

References
CERMAT: Katalog požadavků zkoušek společné části maturitnı́ zkoušky ZKUŠEBNÍ
PŘEDMĚT: INFORMATIKA, vyššı́ úroveň obtı́žnosti. Praha: Centrum pro zjišťovánı́
výsledků vzdělávánı́, 2010.
Gal-Ezer, J., Beeri, C., Harel, D., Yehudai, A., A High-School Program, in Computer Science.
Computer, 1995, 28, 10, pp. 73-80.
Kučera, L.: Kombinatorické algoritmy, SNTL, Praha 1983
Lessner, D.: Computer Science Education on High Schools, in WDS’10 Proceedings of Con-
tributed Papers: Part I - Mathematics and Computer Sciences, Prague, Matfyzpress. 2010,
pp. 110–115.
Töpfer, P.: Algoritmy a programovacı́ techniky. Prometheus, Praha 2007.
VÚP: Rámcový vzdělávacı́ program pro gymnázia. Praha: Výzkumný ústav pedagogický v Praze,
2007.